1. Field of the Invention
The present invention relates to a driving method of a semiconductor device, and more particularly to a driving method of an active matrix type semiconductor device including transistors formed on a semiconductor substrate or an insulating surface.
2. Description of the Related Art
A semiconductor device having an image sensor function is provided with a photoelectric transducer and one or plural transistors for controlling the photoelectric transducer. As the photoelectric transducer, a PN-type photodiode is often used. The other photoelectric transducer includes a PIN-type photodiode, an avalanche diode, an npn embedded diode, a Schottky diode, a phototransistor, a photoconductor for X-rays, and a sensor for infrared rays.
The semiconductor device having the image sensor function is roughly classified into a CCD type and a CMOS type. The semiconductor device of the CMOS type is classified into a passive type in which an amplifying circuit is not mounted, and an active type in which an amplifying circuit is mounted. Since the amplifying circuit has a function to amplify an image signal of an object read by a photoelectric transducer, the influence of noise is low, and accordingly, the active type CMOS semiconductor device in which the amplifying circuit is mounted is often adopted.
In the active type CMOS semiconductor device, an input terminal of the amplifying circuit having high input impedance is connected to an output terminal of the photoelectric transducer. Thus, a region in which information of the object is read does not deteriorate, and the information of the object can be read again and again. This is generally called nondestructive readout.
A method for expanding a dynamic range (light and dark ratio) by using this nondestructive readout and by outputting signals with different storage times has been studied. As an example, as reported in “O. Yadid-Pecht et. al., Proc. SPIE, vol. 2654, pp 82-92, 1996”, a method is studied in which source signal line driving circuits are singly disposed above and below a pixel portion, and signals having different storage times are outputted to each of them. Besides, as another example, as reported in “ISSCC99: p 308:A 640×512 CMOS Image Sensor with Ultra Wide Dynamic Range Floating-Point Pixel-Level ADC”, a method is studied in which a storage time is changed by the power of 2, like T, 2T, 4T, . . . , (2K)×T (here, T denotes a frame period) and read out.
Incidentally, the storage time denotes a time from the initialization of a photoelectric transducer provided in a pixel to the output of a signal from the pixel. In other words, it denotes a time in which a light receiving portion of a photoelectric transducer is irradiated with light and a signal is stored, and is equivalent to a time called an exposure time.
FIG. 3 shows an example of a schematic view of a semiconductor device in which a photoelectric transducer is provided. The semiconductor device of FIG. 3 includes a pixel portion 104, and a source signal line driving circuit 101, a gate signal line driving circuit 102, and a reset signal line driving circuit 103, which are disposed at the periphery of the pixel portion 104. The source signal line driving circuit 101 includes a biasing circuit 101a, a sample hold circuit 101b, a signal output line driving circuit 101c, and a final output amplifying circuit 101d. 
The pixel portion 104 includes a plurality of pixels 100 arranged in a matrix form. In the pixel portion 104, x columns (vertical)×y rows (horizontal) pixels 100 are provided in the matrix form (x and y are natural numbers).
FIG. 4 is a circuit diagram of the pixel 100 provided at an i-th row and a j-th column. Each pixel 100 corresponds to a region surrounded by one of signal output lines (S1 to Sx), one of power supply lines (VB1 to VBx), one of gate signal lines (G1 to Gy), and one of reset signal lines (R1 to Ry). Besides, each pixel 100 includes a switching transistor 112, an amplifying transistor 113, a resetting transistor 114, and a photoelectric transducer 111.
The potential of the photoelectric transducer 111 provided in each pixel 100 is changed by irradiation of light reflected from an object.
When the gate signal line (Gi) is selected in a state where the potential of the photoelectric transducer has been changed by the irradiation of light, the switching transistor 112 connected to the gate signal line (Gi) is turned on, and a signal corresponding to the potential of the photoelectric transducer 111 is outputted to the signal output line (Sj) through the switching transistor 112. Then, the signal outputted to the signal output line (Sj) is outputted to the source signal line driving circuit 101.
Here, a driving method of the semiconductor device having the foregoing structure will be described with reference to FIG. 15. In FIG. 15, the horizontal axis indicates the time. Incidentally, one frame period (F) is a period from a point when a reset signal is applied to a reset signal line R (any one of R1 to Ry) to a point when a reset signal is again applied, and a horizontal scanning period (P) is a period from a point when a signal is applied to a reset signal line R to a point when a signal is applied to a reset signal line R of a next row.
First, a reset signal line (R1) is selected by a reset signal inputted from the reset signal line driving circuit 103 to the reset signal line (R1) of the first row. Incidentally, in the present specification, that the reset signal line is selected means all the resetting transistors 114 connected to the reset signal line are turned on. That is, here, the resetting transistors 114 of all pixels (pixels of the first row) connected to the reset signal line (R1) are turned on. Then, the photoelectric transducers 111 provided in the pixels of the first row are initialized.
Then, at the same time as the termination of the selection of the reset signal line (R1), a reset signal line (R2) of a next row is selected. Next, the resetting transistors 114 of all the pixels connected to the reset signal line (R2) are turned on, and the photoelectric transducers 111 provided in the pixels of the second row are initialized.
In this way, all the reset signal lines (R1 to Ry) are selected in sequence. Then, the photoelectric transducers 111 provided in the pixels 100 connected to the selected reset signal line R is initialized.
Next, signals applied to the gate signal lines (G1 to Gy) will be described. When six horizontal scanning periods (6×P) have passed since the reset signal was inputted to the reset signal line (R1) of the first row, the gate signal line (G1) is selected by a gate signal inputted from the gate signal line driving circuit 102 to the gate signal line (G1). Then, the switching transistors 112 connected to the gate signal line (G1) are turned on, and signals are outputted to the signal output lines (S1 to Sx) by the pixels 100 of the first row. Incidentally, a storage time (L) of the signal outputted by the pixel 100 in this case is the six horizontal scanning periods (6×P).
Next, the gate signal line (G2) of the second row is selected by a gate signal inputted from the gate signal line driving circuit 102 to the gate signal line (G2) of the second row. Then, the switching transistors 112 connected to the gate signal line (G2) are turned on, and signals are outputted to the signal output lines (S1 to Sx) by the pixels 100 of the second row. A storage time (L) of the signal outputted by the pixel 100 in this case is the six horizontal scanning periods (6×P).
In this way, all the gate signal lines (G1 to Gy) are selected in sequence. Then, the signals of the pixels 100 connected to the selected gate signal line (G1 to Gy) are outputted to the signal output lines (S1 to Sx). As is apparent from FIG. 15, when this driving method is used, the storage times (L) of the signals of the pixels 100 outputted by the pixels 100 are identical to one another, and each of them is the six horizontal scanning periods (6×P).
Subsequently, the relation among the timing of the gate signal outputted to the gate signal line (G1 to Gy), the timing of the reset signal outputted to the reset signal line (R1 to Ry), and the potential of the photoelectric transducer 111 provided in the pixel 100 at the i-th row and j-th column will be described with reference to FIG. 16.
First, the reset signal line (Ri) is selected by the reset signal inputted from the reset signal line driving circuit 103 to the reset signal line (Ri). Then, the resetting transistors 114 of all the pixels 100 (pixels 100 of the i-th row) connected to the reset signal line (Ri) are turned on. Then, the photoelectric transducers 111 included in the pixels 100 of the i-th row are initialized.
After the photoelectric transducer 111 is initialized, when the photoelectric transducer 111 is irradiated with light, an electric charge corresponding to the intensity of light is generated in the photoelectric transducer 111. Then, the electric charge charged in the photoelectric transducer 111 is gradually discharged by the reset operation, and the potential of an n-channel side terminal of the photoelectric transducer 111 becomes low.
As shown in FIG. 16, in the case where the photoelectric transducer 111 is irradiated with a bright light, since the amount of discharge is large, the potential of the n-channel side terminal of the photoelectric transducer 111 becomes low. On the other hand, in the case where the photoelectric transducer 111 is irradiated with a dim light, the amount of discharge is small, and the potential of the n-channel side terminal of the photoelectric transducer 111 does not become very low as compared with the case where the bright light is irradiated.
Then, when the six horizontal scanning periods (6×P) have passed since the reset signal was inputted to the reset signal line (Ri), the gate signal line (Gi) is selected by the gate signal inputted from the gate signal line driving circuit 102 to the gate signal line (Gi) of the i-th row. Then, the switching transistor 112 connected to the gate signal line (Gi) is turned on, and the potential of the n-channel side terminal of the photoelectric transducer 111 is read out as a signal. This signal is proportional to the intensity of the light irradiated to the photoelectric transducer 111.
Incidentally, when light is irradiated, the potential of the n-channel side terminal of the photoelectric transducer 111 becomes low, and when a very bright light is irradiated, the potential of the n-channel side terminal becomes as low as the potential of a power supply reference line 121. When the potential becomes as low as the potential of the power supply reference line 121, the potential of the n-channel side terminal becomes constant, and therefore, such a state is called a saturated state.
The photoelectric transducer 111 stores the electric charge generated by the light irradiated in the storage time. Accordingly, when the storage time varies, even if light of the same intensity is irradiated, since the total amount of electric charge generated by the light varies, the value of the signal also varies. For example, in the case where an intense light is irradiated to the photoelectric transducer 111, it is saturated in a short storage time. Even in the case where a feeble light is irradiated to the photoelectric transducer 111, if the storage time is long, it reaches the saturated state sooner or later. That is, the signal is determined by the product of the intensity of the light irradiated to the photoelectric transducer 111 and the storage time.
In FIG. 16, at the point when the gate signal is inputted, although the potential of the photoelectric transducer 111 irradiated with the dim light is slightly lower than that at the point when the reset signal is inputted, it does not yet reach the saturated state.
On the other hand, the photoelectric transducer 111 irradiated with the bright light is already in the saturated state. In this case, a signal outputted from the pixel 100 can not be accurately read. Thus, it is preferable that in the case where the signal of the pixel 100 including the photoelectric transducer 111 irradiated with the bright light is read, the storage time is a little shorter.
When the foregoing driving method of the semiconductor device is used, the storage times (L) of all the signals outputted from the pixel 100 are the six horizontal scanning periods (6×P), and in other words, all the signals outputted from the pixels 100 can be outputted only in the same storage time.
Thus, in the case where the intensity of light irradiated to the pixel 100 is high, since the potential of the photoelectric transducer 111 comes to have the saturated state, information of an object can not be accurately read. In the case where the intensity of light irradiated to the pixel 100 is low, since the change of potential of the photoelectric transducer 111 is faint, signals outputted from the pixel 100 are not very different from one another, and the information of the object can not be accurately read.
When the method reported in “O. Yadid-Pecht et. al., Proc. SPIE, vol. 2654, pp 82-92, 1996” is used, storage times of signals outputted from pixels have only two kinds. Further, since the driving circuits are singly disposed above and below the pixel portion, there is also a defect that the driving circuit portion becomes large.
In the case where the method reported in “ISSCC99: p 308:A 640×512 CMOS Image Sensor with Ultra Wide Dynamic Range Floating-Point Pixel-Level ADC” is used, storage times of signals outputted from pixels are changed like T, 2T, 4T, . . . , (2k)×T. As a result, there is a defect that when k increases, a readout time becomes very long. For example, in the case of k=3 (in the case where the dynamic range is expanded by a factor of 8), it becomes necessary to take a readout time eight times as long as a normal readout time.